@bongos/core@1.19.1053
Cloud Bongos — the AI-first build platform core (GDS + platform surfaces + module system), installed as a versioned dependency (ADR 0108).
AI Security Review
scanned 5d ago · by lpm-firewall-aiSource inspection confirms a consent-gated transcript analysis feature, but its default-yes prompt and claim that raw transcripts stay local do not clearly establish affirmative consent for sending transcript content to the Claude provider. The separate scanner account-action hints resolve to local skill linting and do not support a block.
Decision evidence
public snapshot- Interactive setup says session logs will be read and describes the analysis, but an empty response to its [Y/n] prompt is treated as consent.
- The session-start pipeline reads and spools Claude transcripts when consent is active; analysis is limited to the onboarding window.
- Analysis builds a prompt from the transcript and sends it to the default Claude CLI provider through stdin.
- The resulting summary and friction evidence are posted to the configured instance inbox; the setup disclosure says raw transcripts stay on the machine.
- The flagged skill-lint locations only inspect and report on skill files; they do not perform the indicated account action.
- Both the transcript staging and upload paths check for saved consent and an unexpired onboarding window; non-interactive setup records no consent.
- The inbox receives distilled analysis output rather than the raw transcript, and the API base resolves from the current instance configuration.
Source & flagged code
24 flagged · loading sourcePackage contains a critical-looking secret pattern.
tests/public_mirror_export.mjsView on unpkg · L103AWS access key ID in tests/public_mirror_export.mjs
tests/public_mirror_export.mjsView on unpkg · L103Findings
Affected versions and remediation
This report applies to @bongos/core@1.19.1053.
See all versions of @bongos/core for other recorded verdicts.
- Review the evidence and your use of @bongos/core@1.19.1053 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.